Error-Correcting Codes That Enable Efficient SNARKs

SNARKs built from error-correcting codes have extremely efficient provers, reasonable verifier costs, and plausible post-quantum security. As such, they are widely used in industry and improving their efficiency remains an active area of research. Naturally, both prover and verifier costs depend heavily on the choice of error-correcting code. In this talk, I will give an overview of the families of error-correcting codes that yield efficient SNARKs, and distill the core properties that enable this efficiency. Furthermore, I will discuss some tradeoffs between prover-friendly properties and verifier-friendly properties. Lastly, I will conclude with some open questions.